
NSF Org: |
CCF Division of Computing and Communication Foundations |
Recipient: |
|
Initial Amendment Date: | September 17, 2013 |
Latest Amendment Date: | September 17, 2013 |
Award Number: | 1331692 |
Award Instrument: | Standard Grant |
Program Manager: |
Sankar Basu
sabasu@nsf.gov (703)292-7843 CCF Division of Computing and Communication Foundations CSE Directorate for Computer and Information Science and Engineering |
Start Date: | October 1, 2013 |
End Date: | September 30, 2017 (Estimated) |
Total Intended Award Amount: | $333,000.00 |
Total Awarded Amount to Date: | $333,000.00 |
Funds Obligated to Date: |
|
History of Investigator: |
|
Recipient Sponsored Research Office: |
1608 4TH ST STE 201 BERKELEY CA US 94710-1749 (510)643-3891 |
Sponsor Congressional District: |
|
Primary Place of Performance: |
CA US 94704-5940 |
Primary Place of
Performance Congressional District: |
|
Unique Entity Identifier (UEI): |
|
Parent UEI: |
|
NSF Program(s): |
Information Technology Researc, CyberSEES |
Primary Program Source: |
|
Program Reference Code(s): | |
Program Element Code(s): |
|
Award Agency Code: | 4900 |
Fund Agency Code: | 4900 |
Assistance Listing Number(s): | 47.070 |
ABSTRACT
This project investigates the potential of a voluntary incentive-based demand response program in limiting the operating risks of future sustainable distribution power grids. The approach is to (1) create and assess real-world experiments of coupon incentive-based demand response (CIDR) in representative distribution grids; (2) formulate optimization-oriented CIDR models by taking into account heterogeneous consumers and risk aware load serving entities (LSEs); (3) introduce efficient algorithms for computing risk and welfare outcomes from CIDR; and (4) design and test novel mechanisms tailored for LSEs' risk-averse objectives. The research activities include stochastic modeling of distribution-level grid operations, design of algorithms and incentive mechanisms, and real-world tests of CIDRs in representative distribution grids.
By providing a cyber-enabled sustainable pathway towards deep integration of intelligent decision makers in the smart distribution grid, this project leads toward an overall cost-effective and environmentally benign utilization of a future energy system portfolio. The research program is strongly coupled with an educational effort to train future leaders in the smart grid industry. A new interdisciplinary course on cyber-enabled engineering and economics of sustainable energy systems is offered to students across multiple universities, including Texas A&M and UC Berkeley. As mentors in a NSF REU program at Texas A&M, the team works directly with undergraduate students, in particular under-represented groups, to empirically test and improve the smart phone CIDR application. The team also plans to actively participate in the community outreach activities such as summer open house activities for high school students and teachers.
PUBLICATIONS PRODUCED AS A RESULT OF THIS RESEARCH
Note:
When clicking on a Digital Object Identifier (DOI) number, you will be taken to an external
site maintained by the publisher. Some full text articles may not yet be available without a
charge during the embargo (administrative interval).
Some links on this page may take you to non-federal websites. Their policies may differ from
this site.
PROJECT OUTCOMES REPORT
Disclaimer
This Project Outcomes Report for the General Public is displayed verbatim as submitted by the Principal Investigator (PI) for this award. Any opinions, findings, and conclusions or recommendations expressed in this Report are those of the PI and do not necessarily reflect the views of the National Science Foundation; NSF has not approved or endorsed its content.
Integration of renewable power sources (solar and wind) into the electric grid is difficult because these sources are intermittent and variable. Consequently, these power sources cannot be dispatched in a controlled manner as can thermal power plants. Therefore, there is a mismatch or imbalance between the production of renewable power and the demand for electric power. Traditionally, the balance is reached by dispatching more or less power from reserve thermal generators to match the demand. This approach is adequate so long as the penetration of renewable generation is low, say under 20 percent. But as renewable generation increases (as is happening everywhere) this approach becomes unviable: it is too expensive and increases greenhouse gases. Alternative approaches are needed to mitigate the imbalance.
Our principal idea is to use the inherent flexibility in consumer demand to accommodate the uncertainty in the supply of renewable power, and thereby to reduce as much as possible the need for reserve thermal power. By saying that a consumer’s demand is flexible we mean that there is a set of power consumption profiles each of which meets the demand of a task. The larger is this set the more flexible is the demand, and the more likely that it can be adequately met by a random and interruptible renewable power source. For example, the task of charging an EV can be met by any power supply profile that provides a specified amount of energy between say 8PM and 6AM. As another example, tasks that can provide demand response are those that can be met by reducing their power consumption during the specified critical time interval and perhaps increasing the consumption during another interval. In other words: instead of controlling thermal reserve generators to follow demand, we propose to tailor flexible demand to match variable renewable generation.
Formulated in this way, flexible demand leads to an abstract allocation or scheduling problem: Allocate or schedule a particular (random) power supply profile to meet a collection of flexible demands. In turn this leads first to the question of feasibility (does such an allocation exist), and second to the questions of computation (find such an allocation) and implementation (can the computation be done online).
The second idea is to parameterize significant classes of flexible demand whose allocation problem can be solved and whose solution has an economic interpretation that can be used to incentivize consumers to offer flexible demand. We have thoroughly investigated flexible tasks characterized by demand profiles which need a certain amount of energy but which can be started and stopped at variable times. These tasks are flexible because their execution can be deferred within certain limits. We use the term deferrable demand (DD) to refer to such preemptible demands.
We have investigated several classes of DD services. The basic class consists of tasks that must receive (say) 1kW for h out of T hours (for example, 3 out of 24 hours). The flexibility resides in the fact that the demand is indifferent as to which h hours power is provided to meet the task. The smaller h is, the more flexible is this task. The power can be provided at any hour during [0, T]. There are n tasks characterized by their duration requirements h1, ..., hn. These tasks "arrive" i.e. beome known, at time 0 and they must be completed at some time before T, so their deadline is T. Also given is a power profile p1, ..., pT. This means that pt is the renewable power available in hour t. But renewable power is uncertain. So, unlike information about the tasks, pt becomes known only at the beginning of hour t.
A feasible allocation assigns power A(i, t) = 1 or 0 to task i so that
A(i, 1) + ... + A(i, T) = hi , (task requirements)
A(1, t) + ... + A(n, T) ≤ pt, (power supply constraint)
are met.
Feasibility is then characterized as a set of T linear inequalities involving the {hi} and {pt}. Further an online allocation yields a feasible allocation if one exists. Lastly, there is a market implementation in which DD services, indexed by duration h, are sold at price p(h) that supports a feasible allocation. Moreover, p(h) is increasing in h, which provides an incentive to consumers to select tasks with smaller h (more flexible).
The tasks in the basic model all have the same deadline T. Now suppose that task i is specified by its duration hi and a deadline Ti . A feasible allocation A(i, t) is defined as before except for the additional constraint A(i, t) = 0 for t ≥ Ti , reflecting the deadline.
This is a more difficult problem; nevertheless feasibility is again characterized by linear inequalities that can be interpreted to yield a market equilibrium.
Last Modified: 12/20/2017
Modified by: Pravin P Varaiya
Please report errors in award information by writing to: awardsearch@nsf.gov.